The weak completion semantics and equality

Emmanuelle Anna Dietz Saldanha, Steffen Hölldobler, Sibylle Schwarz, Lim Yohanes Stefanus

Research output: Contribution to journalConference articlepeer-review

2 Citations (Scopus)

Abstract

The weak completion semantics is an integrated and computational cognitive theory which is based on normal logic programs, three-valued ̷Lukasiewicz logic, weak completion, and skeptical abduction. It has been successfully applied – among others – to the suppression task, the selection task, and to human syllogistic reasoning. In order to solve ethical decision problems like – for example – trolley problems, we need to extend the weak completion semantics to deal with actions and causality. To this end we consider normal logic programs and a set E of equations as in the fluent calculus. We formally show that normal logic programs with equality admit a least E-model under the weak completion semantics and that this E-model can be computed as the least fixed point of an associated semantic operator. We show that the operator is not continuous in general, but is continuous if the logic program is a propositional, a finite-ground, or a finite datalog program and the Herbrand E-universe is finite. Finally, we show that the weak completion semantics with equality can solve a variety of ethical decision problems like the bystander case, the footbridge case, and the loop case by computing the least E-model and reasoning with respect to this E-model. The reasoning process involves counterfactuals which is necessary to model the different ethical dilemmas.

Original languageEnglish
Pages (from-to)326-342
Number of pages17
JournalEPiC Series in Computing
Volume57
DOIs
Publication statusPublished - 2018
Event22nd International Conference on Logic for Programming, Artificial Intelligence and Reasoning, LPAR 2018 - Awassa, Ethiopia
Duration: 17 Nov 201821 Nov 2018

Fingerprint Dive into the research topics of 'The weak completion semantics and equality'. Together they form a unique fingerprint.

Cite this